Policy-as-Code and Compliance Automation for Secrets Governance in Agentic AI Workloads in Regulated Industries
DOI:
https://doi.org/10.47941/jts.4121Keywords:
Policy as Code, Secrets Management, Agentic AI, Non-Human Identity, Compliance Automation, Zero Trust, Regulated IndustriesAbstract
Purpose: This study examined the governance gap that opens when autonomous agents acquire, use, and propagate credentials dynamically, and proposed a reference framework, the Agentic Secrets Governance Framework (ASGF), to narrow it.
Methodology: The framework couples a SPIFFE and SPIRE based attestation layer with an Open Policy Agent decision point, a dynamic secrets broker that issues short lived, task scoped credentials, and a compliance control library that expresses obligations drawn from the HIPAA Security Rule, PCI DSS 4.0.1, the EU Digital Operational Resilience Act, the NIST AI Risk Management Framework, and ISO/IEC 42001 as machine evaluable Rego policies. A reference implementation was built and evaluated in a controlled three node Kubernetes testbed. Two workflows, healthcare eligibility determination and card payment fraud triage, were selected by purposive sampling, and ten thousand simulated requests were replayed per configuration.
Findings: Policy enforced, ephemeral credential issuance reduced the configured credential exposure window from months to minutes while adding a combined median of 34.3 milliseconds to the request path. In conformance testing, the policy layer denied the out of scope identity abuse and credential replay cases it was designed to deny, but, as expected, it did not stop goal hijack attacks that stayed within an agent's nominal scope.
Unique Contribution to Theory, Practice and Policy: The study frames secrets governance for agents as a joint security and compliance decision made at the moment of credential issuance. It gives practitioners a reusable control library pattern and issuance protocol, and it suggests that regulators could reduce interpretive risk by publishing machine evaluable reference controls alongside regulatory text. Residual risks from semantic intent verification and cross agent delegation are discussed.
Downloads
References
[1] CrewAI, "2026 State of Agentic AI Survey," Feb. 11, 2026. [Online]. Available: https://www.businesswire.com/news/home/20260211693427/en/
[2] K. Greshake, S. Abdelnabi, S. Mishra, C. Endres, T. Holz, and M. Fritz, "Not what you've signed up for: Compromising real-world LLM-integrated applications with indirect prompt injection," in Proc. 16th ACM Workshop on Artificial Intelligence and Security (AISec), Copenhagen, Denmark, Nov. 2023, pp. 79-90, doi: 10.1145/3605764.3623985.
[3] Q. Zhan, Z. Liang, Z. Ying, and D. Kang, "InjecAgent: Benchmarking indirect prompt injections in tool-integrated large language model agents," in Findings of the Association for Computational Linguistics: ACL 2024, Bangkok, Thailand, Aug. 2024.
[4] E. Debenedetti, J. Zhang, M. Balunović, L. Beurer-Kellner, M. Fischer, and F. Tramèr, "AgentDojo: A dynamic environment to evaluate prompt injection attacks and defenses for LLM agents," in Advances in Neural Information Processing Systems 37 (NeurIPS 2024), Datasets and Benchmarks Track, 2024.
[5] M. Meli, M. R. McNiece, and B. Reaves, "How bad can it git? Characterizing secret leakage in public GitHub repositories," in Proc. Network and Distributed System Security Symposium (NDSS), San Diego, CA, USA, Feb. 2019, doi: 10.14722/ndss.2019.23418.
[6] A. Nabiullina and C. Winqwist, "The State of Secrets Sprawl 2026," GitGuardian, Mar. 17, 2026. [Online]. Available: https://www.gitguardian.com/state-of-secrets-sprawl-report-2026
[7] Cloud Security Alliance, "The Non-Human Identity Governance Vacuum: Agentic AI Governance," CSA Whitepaper, May 20, 2026. [Online]. Available: https://labs.cloudsecurityalliance.org/research/csa-whitepaper-nonhuman-identity-agentic-ai-governance-v1-cs/
[8] T. Otsuka, K. Toyoda, and A. Leung, "AI Identity: Standards, Gaps, and Research Directions for AI Agents," arXiv:2604.23280, Apr. 28, 2026.
[9] HashiCorp, "Advancing AI Agent Security in Vault," HashiCorp Blog, Jun. 23, 2026. [Online]. Available: https://www.hashicorp.com/en/blog/advancing-ai-agent-security-in-vault
[10] HashiCorp, "SPIFFE: Securing the Identity of Agentic AI and Non-Human Actors," HashiCorp Blog, 2026. [Online]. Available: https://www.hashicorp.com/en/blog/spiffe-securing-the-identity-of-agentic-ai-and-non-human-actors
[11] Y. Ruan, H. Dong, A. Wang, S. Pitis, Y. Zhou, J. Ba, Y. Dubois, C. J. Maddison, and T. Hashimoto, "Identifying the risks of LM agents with an LM-emulated sandbox," in Proc. 12th Int. Conf. Learning Representations (ICLR), Vienna, Austria, 2024.
[12] OWASP Foundation, "OWASP Top 10 for Agentic Applications 2026," Generative AI Security Project, Dec. 9, 2025. [Online]. Available: https://genai.owasp.org/resource/owasp-top-10-for-agentic-applications-for-2026/
[13] Cloud Security Alliance, "MITRE ATT&CK and ATLAS Agentic Gap Analysis: Techniques Unique to Autonomous Agent Control Planes," CSA Research Note, Mar. 27, 2026. [Online]. Available: https://labs.cloudsecurityalliance.org/agentic/csa-research-note-atlas-agentic-gap-analysis-20260327/
[14] S. K. Basak, L. Neil, B. Reaves, and L. Williams, "What are the practices for secret management in software artifacts?," in Proc. IEEE Secure Development Conference (SecDev), 2022, pp. 69-76.
[15] T. Lodderstedt, J. Richer, and B. Campbell, "OAuth 2.0 Rich Authorization Requests," RFC 9396, Internet Engineering Task Force, May 2023.
[16] HashiCorp, "Manage Vault policies with Sentinel," Vault Enterprise Documentation, 2026. [Online]. Available: https://developer.hashicorp.com/vault/docs/enterprise/sentinel
[17] HashiCorp, "SPIFFE auth method," Vault Documentation, 2026. [Online]. Available: https://developer.hashicorp.com/vault/docs/auth/spiffe
[18] SPIFFE Project, "SPIRE Concepts," Cloud Native Computing Foundation, 2026. [Online]. Available: https://spiffe.io/docs/latest/spire-about/spire-concepts/
[19] S. Rose, O. Borchert, S. Mitchell, and S. Connelly, "Zero Trust Architecture," NIST Special Publication 800-207, National Institute of Standards and Technology, Aug. 2020, doi: 10.6028/NIST.SP.800-207.
[20] S. T. Avirneni, "Establishing Workload Identity for Zero Trust CI/CD: From Secrets to SPIFFE-Based Authentication," arXiv:2504.14760, Apr. 2025.
[21] J. W. Cutler et al., "Cedar: A new language for expressive, fast, safe, and analyzable authorization," Proc. ACM Program. Lang., vol. 8, no. OOPSLA1, Apr. 2024, doi: 10.1145/3649835.
[22] M. Jones, A. Nadalin, B. Campbell, J. Bradley, and C. Mortimore, "OAuth 2.0 Token Exchange," RFC 8693, Internet Engineering Task Force, Jan. 2020.
[23] T. South, S. Marro, T. Hardjono, R. Mahari, C. D. Whitney, D. Greenwood, A. Chan, and A. Pentland, "Authenticated delegation and authorized AI agents," arXiv:2501.09674, Jan. 2025.
[24] G. Syros, A. Suri, J. Ginesin, C. Nita-Rotaru, and A. Oprea, "SAGA: A security architecture for governing AI agentic systems," in Proc. Network and Distributed System Security Symposium (NDSS), San Diego, CA, USA, 2026; arXiv:2504.21034.
[25] Microsoft Security, "Shai-Hulud 2.0: Guidance for Detecting, Investigating, and Defending Against the Supply Chain Attack," Microsoft Security Blog, Dec. 9, 2025.
[26] Open Policy Agent, "OPA Documentation," Cloud Native Computing Foundation, 2026. [Online]. Available: https://www.openpolicyagent.org/docs
[27] Open Policy Agent, "Gatekeeper: Policy controller for Kubernetes" and "Conftest," project documentation, 2026. [Online]. Available: https://open-policy-agent.github.io/gatekeeper/ and https://www.conftest.dev
[28] M. Kellogg, M. Schäf, S. Tasiran, and M. D. Ernst, "Continuous compliance," in Proc. 35th IEEE/ACM Int. Conf. Automated Software Engineering (ASE), 2020, pp. 511-523, doi: 10.1145/3324884.3416593.
[29] M. A. Varma, "Compliance-as-Code for Continuous PCI DSS 4.0 Validation in Cloud-Native Financial Systems: Principles, Patterns, and Industrial Experience," International Journal of Intelligent Systems and Applications in Engineering, vol. 14, no. 1s, May 2026.
[30] U.S. Department of Health and Human Services, Office for Civil Rights, "Notice of Proposed Rulemaking: HIPAA Security Rule," Federal Register, Jan. 2025.
[31] PCI Security Standards Council, "Payment Card Industry Data Security Standard, Requirements and Testing Procedures, Version 4.0.1," 2024.
[32] European Parliament and Council of the European Union, "Regulation (EU) 2022/2554 on Digital Operational Resilience for the Financial Sector (DORA)," in force Jan. 17, 2025.
[33] National Institute of Standards and Technology, "Artificial Intelligence Risk Management Framework (AI RMF 1.0)" and "Generative Artificial Intelligence Profile (NIST AI 600-1)," NIST, 2023-2024.
[34] International Organization for Standardization, "ISO/IEC 42001:2023, Information Technology, Artificial Intelligence Management System," Dec. 2023.
[35] Google Cloud, "Workload Identity Federation," Cloud IAM Documentation, 2026. [Online]. Available: https://cloud.google.com/iam/docs/workload-identity-federation
[36] Amazon Web Services, "What is AWS Identity and Access Management Roles Anywhere?," IAM Roles Anywhere User Guide, 2026. [Online]. Available: https://docs.aws.amazon.com/rolesanywhere/latest/userguide/introduction.html
[37] Microsoft, "What are workload identities?," Microsoft Entra Workload ID Documentation, 2026. [Online]. Available: https://learn.microsoft.com/en-us/entra/workload-id/workload-identities-overview
[38] Microsoft, "Microsoft extends Zero Trust to secure the agentic workforce," Microsoft Security Blog, May 19, 2025. [Online]. Available: https://www.microsoft.com/en-us/security/blog/2025/05/19/microsoft-extends-zero-trust-to-secure-the-agentic-workforce/
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Suhas Madupuri, Rajesh Thondamanati

This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution (CC-BY) 4.0 License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.